In recent years, generative AI, which can respond interactively to text-based questions, has been spreading rapidly throughout society. Tasks such as proofreading documents or translation are now performed by AI on a daily basis.
Originally, artificial intelligence (AI) referred to technologies designed to enable machines to perform intellectual tasks once thought possible only for humans, and research in this field has been pursued steadily for more than seventy years. Over the past decade, however, AI has made dramatic progress, bringing it much closer to human capabilities.
In this talk I will discuss, from my standpoint as an algorithm specialist, what kinds of computation today’s AI performs well, and what kinds remain difficult.